An empirical note on weather effects in the Australian stock market Author Worthington, Andrew Published 2009 Journal Title Economic Papers DOI https://doi.org/10.1111/j.1759-3441.2009.00014.x Copyright Statement © 2009 The Economic Society of Australia. Published by Blackwell Publishing. This is the pre-peer reviewed version of the following article: Economic Papers Vol. 28, No. 2, June, 2009, 148–154 which has been published in final form at http://dx.doi.org/10.1111/ j.1759-3441.2009.00014.x Downloaded from http://hdl.handle.net/10072/30698 Link to published version http://www.interscience.wiley.com/jpages/0812-0439 Griffith Research Online https://research-repository.griffith.edu.au AN EMPIRICAL NOTE ON WEATHER EFFECTS IN THE AUSTRALIAN STOCK MARKET ABSTRACT The behavioural finance literature posits a link between the weather and equity markets via investor moods. This paper examines the impact of weather on the Australian stock market over the period 1958 to 2005. A regression-based approach is employed where daily market returns on the Australian Securities Exchange’s All Ordinaries price index are regressed against eight daily weather observations (precipitation, evaporation, relative humidity, maximum and minimum temperature, hours of bright sunshine, and the speed and direction of the maximum wind gust) at Sydney’s Observatory Hill and Airport meteorological stations. Consistent with studies elsewhere including the Australian market, the results indicate that the weather has absolutely no influence on market returns. Some directions for future research that may help address some of the deficiencies found in this intriguing body of work are provided. JEL CODE: C32; G12; G14 KEY WORDS: Weather effects; Market efficiency; Investor moods 1. INTRODUCTION A well-established and diverse literature, primarily in the field of psychology, has investigated the premise that “…weather variables affect an individual’s emotional state or mood, which creates a predisposition to engage in particular behaviours” (Howarth and Hoffman, 1984, p. 15). An almost equally well-known theoretical extension has examined the importance of moods in economic decision-making (see, e.g., Elster, 1998; Loewenstein, 2000; Romer, 2000; Loewenstein et al., 2001; Hanock, 2002). One strand of this work suggests that the moods of investors may affect equity prices if investor’s subjective preferences (including the level of risk aversion and their judgement of the appropriate discount rate) fluctuate over time, if the effects of these fluctuations are widely and uniformly experienced, and if investors do not realise their decisions are influenced by fluctuations in their moods (Mehra and Sah, 2002). Lucey and Dowling (2005) provide a useful survey of the role of moods in investor decision-making. Critically, this complements very recent empirical work that has sought to investigate whether the positive (negative) moods induced by good (bad) weather cause a mood misattribution that results in marginal investors pricing stocks more optimistically (pessimistically) (see, e.g., Saunders, 1993; Keef and Roush, 2003; Hirshshleifer and Shumway, 2003; Pardo and Valor, 2003; 2 Garrett et al., 2003; Loughran and Schultz, 2004; Goetzmann and Zhu, 2005; Cao and Wei, 2005). The purpose of this note is to add to this intriguing body of work the results of an analysis of weather and its impact on the Australian equity market. Although the Australian market has been partially addressed in studies of international weather effects by Hirshleifer and Shumway (2003), Cao and Wei (2005) and Garrett et al. (2005), a comprehensive analysis remains, as yet, undone. Using daily data over a long time period, and a wide range of weather indicators as proxies for mood factors, this paper confirms that there is no evidence to support the presence of a weather effect in Australian stock returns. The remainder of the note is organised as follows. Section 2 briefly reviews the literature on weather and its influence on investor decision making. Section 3 explains the empirical methodology and data employed in the study and Section 4 provides a descriptive analysis. The empirical findings are presented and analysed in Section 5. The paper ends with a brief conclusion. 2. LITERATURE REVIEW The seminal paper on the relationship between weather-induced moods and equity returns is Saunders (1993). Using daily returns from the Dow-Jones Industrial Average from 1927 to 1989 and the New York Stock Exchange and American Stock Exchange from 1962 to 1989, Saunders (1993) examined whether any systematic variation in these markets could be associated with local weather patterns (as observed at New York Central Park from 1927 to 1960 and New York LaGuardia Field from 1961 to 1989). Weather was defined in the form of cloud cover. The results indicated that when cloud cover was 100 percent, returns were below average, but when cloud cover was below 20 percent, returns were above average. Following these remarkable and rather well-publicised findings (see, e.g., Stecklow, 1993; Koretz, 1994) a number of other studies also examined the relationship between weather and stock returns. Trombley (1997, p. 18) reexamined the Dow-Jones Industrial Average over the period 1927 to 1989 and concluded that “…the relationship between security returns and Wall Street weather is neither as clear nor as strong as Saunders (1993) suggests. There is no difference between returns on clear sunny days and on cloudy or rainy days”. Krämer and Runde (1997) also investigated cloud cover, though in the context of the German stock index (DAX) and with three additional weather indicators observed at Frankfurt – humidity, 3 atmospheric pressure and rainfall. Krämer and Runde (1997) also concluded that no systematic relationship seemed to exist. Similar results were quickly confirmed in Turkey (Tufan and Hamarat, 2004), New Zealand (Keef and Roush, 2003) and Spain (Pardo and Valor, 2003). However, Keef and Roush (2003) did find some evidence of a negative temperature and wind effect (as measured at Wellington) on New Zealand stock returns, afterwards a positive sunshine and negative wind effect (Keef and Roush, 2005), and later still a latitude-dependent cloud cover effect (Keef and Roush, 2007b). Dowling and Lucey (2005) also examined a range of weather indicators (as measured at Dublin Airport) for the Irish Stock Exchange, including cloud, precipitation, humidity and the presence of geomagnetic storms, concluding that rain was a minor but significant market influence. Internationally, Hirshleifer and Shumway (2003) considered the relationship between cloud cover and equity returns in 26 markets. As with Saunders (1993), a negative relationship between cloud cover and returns was found, but only in three cases of cloud cover (Milan, Rio de Janeiro and Vienna) and two cases of precipitation (Brussels and New York) were they significant. Cao and Wei (2005) reexamined the international weather effect, though with temperature, concluding a negative correlation between temperature and returns. However, the effect was not ubiquitous. Finally, Kamstra et al. (2003) and later Garrett et al. (2005) investigated the role of seasonal affective disorder (clinical depression associated with shorter winter days) in influencing market returns. These studies built on earlier work by Kamstra et al. (2000) of a daylight saving effect (for rebuttals, see Pinegar, 2002; Worthington, 2003). Kamstra et al. (2003, p. 14) concluded that the results “…were consistent with a [seasonal affective disorder]-induced pattern in returns as depressed and risk-averse investors shun risky assets in the fall and resume their risky holdings in the winter, leading to returns in the fall which are lower than average and returns following the longest night of the year which are higher than average”. Garrett et al. (2005) later found that the seasonal affective disorder effect arose due to the heightened risk aversion that came with seasonal depression, as reflected by a changing risk premium. 3. EMPIRICAL METHODOLOGY The approach used to test for weather effects in Australian stock returns is regression-based 4 where market returns are regressed against a set of weather observations. The following model is specified: N Rt = β0 + ∑βiWit +ε t (1) i=1 where Rt represents the market return on day t, Wit are observations for the ith weather-related variable on day t (where i = 1,2…N), β are parameters to be estimated and ε is the error term. The market data employed are closing prices from the Australian Securities Exchange (ASX) and a predecessor (Sydney stock exchange) from Monday 6 January 1958 to Friday 30 December 2005. This sample encompasses 12,067 trading days and represents the most complete set of daily data available. The capitalization-weighted All Ordinaries Price Index is used. The index covers about 92 percent of domestic companies by market value. To be included in the index, stocks must have an aggregate market value of at least 0.02 percent of all domestic equities, and maintain an average turnover in excess of 0.5 percent of quoted shares each month. The long-term index includes base recalculations by Global Financial Data (2006). A series of market returns are then calculated as Rt = 100ln(Pt Pt−1 ) where Pt is the index at the end of day t. The weather data used in the study are sourced from the NSW Climate and Consultancy Section of the Australian Bureau of Meteorology (2006). The weather stations selected for this analysis (latitude and longitude in brackets) are #066062 Sydney Observatory Hill (– 33.8607, 151.2050) and #066037 Sydney Airport (–33.9411, 151.1725). The former has operated continuously since July 1858 and the latter since September 1929. Sydney is selected over other possible market locations (Melbourne, Brisbane, Adelaide and Perth) because the early market data is largely compiled from old Sydney stock exchange indices. Moreover, Sydney is well accepted as Australia’s financial capital, with existing international work on the weather effect, including Hirshleifer and Shumway (2003), Cao and Wei (2005) and Garrett et al.